The Polar Climate Predictability Initiative (PCPI) is an initiative of the World Climate Research Programme (WCRP), whose goal is to improve the understanding of the predictability of climate and the effect of human activities on climate. The PCPI has a focus on polar regions and their role in the global climate system and aims to improve predictability of the climate system on all time scales by improving our understanding of the underlying physical mechanisms and their representation in climate models.

The PCPI will accomplish this task by coordinating the efforts of the international science community, bringing together the different elements of the WCRP, and working closely with other international agencies such as the World Weather Research Programme’s Polar Prediction Project (WWRP – PPP). The PCPI is addressing the seasonal to multi-decadal component of the GIPPS (Globally Integrated Polar Prediction System) of the World Meteorological Organization, in close co-ordination with the WWRP PPP, which is addressing the shorter time scales. The focus here is not on prediction of the climate system, but instead on finding elements of the climate system that contribute to predictability, and how these processes may be improved in models.

The PCPI complements existing efforts by bringing together expertise on the modeling aspects of the climate. It is an initiative of the WCRP under the Grand Challenge “Cryosphere in a Changing Climate.”

Co-leads: Julie Jones (University of Sheffield, UK; link to CliC) and Marilyn Raphael (UCLA, USA; link to SPARC)

The PCPI is supported by the Climate and Cryosphere Project (CliC) and Atmospheric Processes And their Role in Climate (APARC), two core projects of the WCRP.

Science Background and the Polar Climate Predictability

Arctic climate is responding more rapidly to increasing greenhouse gases and other climate forcings than is the tropics. While Arctic sea-ice extent has been declining over the last few decades, Antarctic sea-ice extent has been increasing. Climate models can simulate the decline in Arctic sea-ice extent but generally do not capture the recent rapid changes in the sea-ice minimum, and they cannot simulate the increase in the Antarctic. Understanding why the poles are more sensitive to changes in climate, and why they respond differently under the same climate forcing is essential to improving future climate projections.

Because polar climate is highly variable, predicting its evolution on seasonal and interannual time scales is important. Tropical climate phenomena, such as the El Nino Southern Oscillation (ENSO) and the Quasi-Biennial Oscillation (QBO), have been studied as possible sources of predictability in the atmosphere. Sea ice and snow cover, the state of the stratosphere (including the ozone hole in the southern hemisphere), and the southern oceans have also been identified as potential sources of predictability in the polar regions. These aspects of the climate system are associated with long-time scales and provide boundary conditions for the troposphere.

Climate at the Poles

The September Arctic sea ice extent in 2012 was at a record minimum. In contrast to the Arctic, Antarctic sea ice extent has been increasing and recorded a record maximum in September 2012 (when Antarctic sea ice is at its seasonal maximum). The growth in Antarctic sea ice is much smaller than the sea ice loss in the Arctic.

Images showing the decline in Arctic sea ice extent (left) and the increase in Antarctic sea ice (right). The median extent is shown by the red and yellow outlines. Credit: NASA/Goddard Space Flight Center Scientific Visualization Studio and NASA Earth Observatory/Jesse Allen

Poles apart

The Northern and Southern polar regions have very different geographies. The Arctic is covered by frozen ocean surrounded by large land masses, and the Antarctic is covered by frozen land surrounded by ocean.

The Arctic Ocean is surrounded by North America, Greenland, and Eurasia. These large landmasses trap most of the sea ice, which builds up and retreats with each yearly freeze-and-melt cycle. But a large fraction of the older, thicker Arctic sea ice has disappeared over the last three decades. The shrinking summer ice cover has exposed dark ocean water that absorbs sunlight and warms up, leading to more ice loss. On the opposite side of the planet, Antarctica is a continent circled by open waters that let sea ice expand during the winter but also offer less shelter during the melt season. Most of the Southern Ocean’s frozen cover grows and retreats every year, leading to little perennial sea ice in Antarctica.

Sources of Predictability

Sea ice and snow cover

Arctic sea ice loss
Arctic September sea ice extent (x106 km2) from observations (thick red line) and 13 IPCC AR4 climate models, together with the multi-model ensemble mean (solid black line) and standard deviation (dotted black line). Models with more than one ensemble member are indicated with an asterisk.

The September Arctic sea ice extent in 2012 was at a record minimum. While current climate models show declining Arctic sea ice over the last half century, they are not able to replicate the rapid pace of the decline, which means that projections of sea ice loss are not accurate. Despite this, current estimates based on the models and statistical methods used to get at the most likely scenario of sea ice loss, predict that September Arctic sea ice will most likely disappear between 2066 and 2085. Other estimates that take into account sea ice volume (rather than extent) project that summer sea ice could vanish as soon as 2059.

September sea-ice cover in the Arctic Ocean projected to vanish by 2100, Boe et al., 2009: Nature Geoscience, 2, 341 – 343 (2009), doi:10.1038/ngeo467

Arctic Sea Ice Decline: Faster than Forecast, Stroeve, et al., 2007: Geophys. Res. Lett., 34, L09501, doi: 10.1029/2007GL029703.

Eurasian snow cover

[Coming Soon]

Stratosphere and the ozone hole

The stratosphere is usually thought of as responding to the troposphere, with disturbances from the troposphere changing the circulation and transport of chemical species. However, recent studies have shown that the stratosphere can have a significant effect on the troposphere.

The ozone hole and tropospheric winds

Photo: Trends in December-to-February (DJF) zonal-mean zonal wind. The multimodel mean trends between 2001 and 2050 are shown for chemistry-climate models (A), climate models (B), models with prescribed ozone recovery (C), and models with no ozone recovery (D). Deceleration and acceleration are indicated with blue and red colors, respectively. Superimposed black solid lines are DJF zonal-mean zonal wind averaged from 2001 to 2010.

Many climate models prescribe ozone levels, rather than model them, in order to make computations more efficient and faster. However, it is clear that models that do not properly prescribe ozone (i.e., the decrease in ozone due to the anthropogenic release of ozone destroying substances, and the predicted recovery of the ozone layer) do not correctely predict the change in the tropospheric jet location in the Southern hemisphere. Even those models that do correctly prescribe ozone but do not model its behaviour, get a smaller shift in the jet than full chemistry-climate models. The figure shows that the ozone hole in the Southern summertime stratosphere has a clear impact on the jet location, and accurate prediction of the tropospheric winds will require a good representation of the ozone hole in climate models.

The Impact of Stratospheric Ozone Recovery on the Southern Hemisphere Westerly Jet, Son et al.,Science 13 June 2008: 320, pp. 1486-1489, DOI: 10.1126/science.1155939

The stratosphere as a source of seasonal predictability

Other studies have shown that improvement in the representation of the stratosphere in weather prediction models gives improvement in forecast skill scores. Seasonal forecasting in particular is becoming more common and providing socio-economic benefits. However, such forecasts are often inaccurate. Seasonal forecasting with models that have an accurate representation of the stratosphere is of potential benefit since long-lived circulation anomalies in the stratosphere can affect the troposphere on long time scales. Shown in the figure are the results from forcasts from a model that does not have an accurate representation of the state of the stratosphere after a Stratosphere Sudden Warming (blue), and a model that does (pink). The skill scores for regional temperatures, precipitation, and sea level-pressure are clearer better (skill score of 1.0 is a perfect forecast) when the stratosphere is accurately represented.

The skill score (CSS) of the (b) surface temperature over northern Russia and eastern Canada, precipitation difference between the high latitude and mid-latitude northern Atlantic, and the CSS averaged over 20−90N for sea-level pressure (b), surface temperature (land alone; c) and precipitation (d). Pink bars represent the forecast with good representation of an SSW, and blue bars are for one without a representation of an SSW for a forecast range of 16−60 days. The thin lines represent the 95% confidence interval.

Enhanced seasonal forecast skill following stratospheric sudden warmings, Sigmond, et al. 2013: Nature Geoscience, 6, 98-102, doi:10.1038/ngeo1698.

Stratosphere-troposphere coupling
Top: Time−height development of the NAM index; white contours indicate NAM values of 1 and 2. Horizontal time axis indicates the lead or lag (in days) with respect to the date of the events. The events are determined by the dates on which the NAM at 10 hPa exceeds +2.5. Bottom: Associated (red) zonal wind stress and (black) SST anomalies over the North Atlantic study region; numbers at the upper right are averages over days 0−60.

The stratosphere is connected to tropospheric weather and climate. In particular, extreme stratospheric circulation events are known to exert a dynamical feedback on the troposphere1. However, it is unclear whether the state of the stratosphere also affects the ocean and its circulation. A co-variability of decadal stratospheric flow variations and conditions in the North Atlantic Ocean has been suggested, but such findings are based on short simulations with only one climate model. Here we assess ocean reanalysis data and find that, over the previous 30 years, the stratosphere and the Atlantic thermohaline circulation experienced low-frequency variations that were similar to each other. Using climate models, we demonstrate that this similarity is consistent with the hypothesis that variations in the sequence of stratospheric circulation anomalies, combined with the persistence of individual anomalies, significantly affect the North Atlantic Ocean. Our analyses identify a previously unknown source for decadal climate variability and suggest that simulations of deep layers of the atmosphere and the ocean are needed for realistic predictions of climate.

A stratospheric connection to Atlantic climate variability, Reichler, et al., 2012: Nature Geoscience, 5, 783-787, doi:10.1038/ngeo1586

Southern Ocean

[Coming Soon]

Climate variability and extremes

[Coming Soon]